Spatially autocorrelated sampling falsely inflates measures of accuracy for presence‐only niche models

By S. Veloz. 2009

641 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Nature and Landscape Conservation

View Paper: https://www.semanticscholar.org/paper/962643ef679dc2e2d2faaabf1ce1c715444163d0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Ecological Modeling

In Ecology

In Nature and Landscape Conservation

Citations in the corpus

Using species distribution modelling to disentangle realised versus potential distributions for rare species conservation (2013, background): “Spatial autocorrelation (SAC) needs to be taken into account in order to avoid inflation of accuracy measures (Veloz, 2009).”
Mammal defaunation as surrogate of trophic cascades in a biodiversity hotspot (2013, background): “These values can be interpreted as the probability of presence of suitable environmental conditions for the target species (Veloz, 2009).”
spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models (2015, methodology): “Veloz (2009) provided one method for determining Johnson, D. S. 1982.”
Conservation biogeography – foundations, concepts and challenges (2010, background): “…our understanding of current distributions, we also need to deploy similar sensitivity analyses to a wide range of issues in modelling future processes and patterns of diversity change (e.g. Veloz, 2009; Willems & Hill, 2009; Feeley & Silman, 2010; Roura-Pascual et al., 2010; Smolik et al., 2010).”
Estimating optimal complexity for ecological niche models: A jackknife approach for species with small sample sizes (2013, methodology): “Finally, other than removing duplicate localities falling into the same map pixel, no spatial filtering was employed (in contrast to the present study; see Section 2.2), likely overestimating model performance due to spatial autocorrelation (Veloz, 2009), again favoring complex models.”
Paintings predict the distribution of species, or the challenge of selecting environmental predictors and evaluation statistics (2018, background): “This approach assumes that training and testing data are spatially independent, an assumption rarely fully met in practice, especially when occurrences are randomly partitioned (Veloz, 2009).”
Spatial filtering to reduce sampling bias can improve the performance of ecological niche models (2014, methodology): “Several studies have used filtering (=thinning) techniques (Anderson and Raza, 2010; Carroll, 2010; Pearson et al., 2007; Veloz, 2009) to reduce the effects of sampling”
What do we gain from simplicity versus complexity in species distribution models (2014, background): “Recent studies have emphasized that care should also be taken in how data is partitioned into training, evaluation and test data, in particular to control for spatial autocorrelation (Latimer et al. 2006, Dormann et al. 2007, Veloz 2009, Hijmans 2012; see below for more details).”
SDMtoolbox: a python‐based GIS toolkit for landscape genetic, biogeographic and species distribution model analyses (2014, background): “When spatial clusters of localities exist, often models are over-ﬁt towards environmental biases (reducing the model’s ability to predict spatially independent data) and model performance values are inﬂated (Veloz 2009; Hijmans 2012; Boria et al. 2014).”
Residual spatial autocorrelation in macroecological and biogeographical modeling: a review (2019, background): “2012 Journal of Biogeography Elaborate Multi-scale SAC & Invasive forest pathogen distribution 27 Veloz 2009 Journal of Biogeography Elaborate Niche modeling and plant species distribution 28 Wu et al. 2013 Applied Geography Elaborate Model comparison and occurrence of cloud cover 29 Siesa et al.…”
Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure (2017, background): “…on random data splits (all of which will be largely consistent in underlying structure) as well as the various parametric modelling options (Table 1) will tend to underestimate predictive error (Araújo et al. 2005, Veloz 2009, Bahn and McGill 2013), leading to false confidence in model predictions.”
The importance of correcting for sampling bias in MaxEnt species distribution models (2013, background): “spatial autocorrelation [SAC]) and affects model quality by inflating model accuracy (Veloz, 2009), leading to type I errors (Dormann et al.”